Microsoft plans science of appliance for next SQL Server Start your clocks and count the delays: Microsoft has named the first half of 2010 as the window for the next version of SQL Server - codenamed Kilimanjaro. Microsoft has also said it's working on an server appliance with hardware giants Bull, Dell, EMC, Hewlett-Packard, and Unisys that'll use a subset of technology in …
Maybe this makes sense but not to me. I'm going to stick my neck out and look stupid, but I have to ask -
1st para, setting the scene
2nd para talks about a "server appliance" - what does it do? Apparently it'll "use a subset of technology in Kilimanjaro" but what subset? A quick scan of the datallegro website just says how marvellous it is with its 'comprehensive, tightly integrated functionality' et bleedin cetera - so what is it? Likewise zoomix, with its 'Automating data enterprise excellence' (I think I'm going to be sick - sorry, review my internal dynamics using a paradigm of self-inversion). And then there's the "large scale and detailed analysis" - of what? how? why? what need does it fill?
3rd para tells us that stuff is happening between MS HP & oracle and talks about a "massive, parallel, high-speed server". Meaningless jabber (why report it?), but then it mentions a fact - Linux. So, Oracle didn't run on linux but will now? what's it saying? And the last bit suggests oracle will run on windows - I thought it did already.
para 4 is just boilerplate MS vapourware announcements. After reading it I don't even know when this thing I don't understand might (but probably won't) be delivered.
para 5 says that "the new version of SQL Server will be focused on business intelligence", but wasn't that the case with 2005 & 2008? (may be wrong about 08) but what BI improvements will it bring? And then there's ""Kilimanjaro will provide "self-service reporting" and "self-service analysis" through something called Project Gemini."" So what is gemini and what is self-service reporting/analysis and how will it persuade me to buy into it? I'm floundering.
Para 6 is more of the same empty guff as para 5
Para 7 is more of the bigger/faster/throbbier with its terabytes and huge numbers of concurrent users for some undefined value of concurrent, a rather multidimensional attribute IMO.
Wind up with two more paras of empty blather except perhaps for "new data warehouse reference configurations" which might mean something but I don't know what, and I'm done.
Now I don't mind looking stupid if I learn something from it, and I freely own that this is not the area of DBs in which my experience lives, so perhaps Gavin Clarke or someone could tell me in plain language, *WHAT IT WAS I JUST READ*??
That's when non-technical people use a point-and-drool interface to kludge together reports that slow the server to a crawl, then pound F5 in their browsers in the futile hope that they will find enlightenment.
To me, it still looks like an old, platform-restricted version of Sybase Adaptive Server. Microsoft seem to be repeating a (recently frequent) mistake of trying to tackle Linux in its home ground, with appliances this time.
I always thought a large part of the argument for appliances was near-zero administration - a lot of companies use one or two Linux based appliances because their IT staff are largely Windows-centric with minimal if any Linux experience and skills. Since the (only possible) argument for using a GUI on a server is ease of administration and existing Windows skills, what's the point of a Windows appliance?
I dread to think what a Windows appliance would run like after a few months of zero admin anyway. This is almost as crazy as running Windows on a supercomputer.
CERN is using oracle to manage the huge amounts of data generated by the LHC.
Given that CERN has also created the largest federated grid parallel supper-computer this little planet has seen.. you can be fairly sure Infiniband is in the mix.
You can all be fairly sure that the Oracle Exadata appliance was build for CERN.. so get ready for the strapline.
The irony is that Microsoft was leading in this area with Federated SQL/Server over VIA.. but dropped the ball.
I think I have finally figured out what Sun are trying to do; and if I'm right, it's going to be backs-to-the-wall time for Microsoft soon.
Acquring StarOffice and opening up the code seemed a strange move (but anyone who saw the schoolboy errors in the OO.org 1.x codebase that StarDivision thought nobody would ever see, will tell you: it got the code cleaned up). Solaris and Java were more scene-setting. Acquiring MySQL at the same time as the infamous blog backend stopped being a toy and started looking like a real database was moving the Queen into play. Sun have a winner on their hands: the ability to combine the popular Java language with MySQL (can anyone spell stored procedures?) can only increase its power.
Sun have a long-term strategy. They are beginning, slowly but surely, to groom knowledgeable IT people to expect to receive the Source Code when they choose well-known, high-end products. More than one MS-hating school-leaver who ends up getting a job in a small-to-medium business and weaning them off Microsoft and onto Ubuntu will end up in a position to specify Sun hardware and software. And Sun know that. Gates and Ballmer were the reason these people walked out on MS; but The Source was the reason they stayed.
Vista was already seen by everyone outside Microsoft as Pawn to King's Knight Four. SQL Server 2010 could just be the Pawn to King's Bishop Three that follows.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
